Privacy preserving data publishing: a coalitional game theory perspective

Abstract

k-anonymity is one of the most popular conventional techniques for protecting the privacy of an individual. In this process, the following limitations are observed: 1 The anonymisation is done based on an assumed value of k; 2 the information loss can be found only after the anonymisation is done; 3 if the information loss is found to be more than the affordable loss then another k is to be considered and the whole process has to be repeated. This paper discusses a novel approach using coalitional game theory CGT to overcome the limitations of k-anonymity. The approach helps fix up the privacy levels based on the information loss. To achieve anonymity, we establish coalitions between the tuples based on their payoffs which are assigned using concept hierarchy tree CHT of quasi identifiers QID. In the process, an attempt has been made to obtain a relation between k and number of distinct tuples with respect to QID set. This helps to find the boundaries of k. The experimental results showing the practicality and scalability are presented.